Mike Dockery & Sherry BawaBankwest Curtin Economics Centre/School of Economics and Finance

Curtin Business School

Labour market implications of promotingwomen’s participation in STEM in Australia

The Policy Context• The “knowledge economy”

• Need to increase proportion of workforce with tertiary qualifications (Bradley Review)

• STEM skills critical to innovation and economic growth (Australian Industry Group, Office of the Chief Scientist, PwC)“Jobs of the future”International comparisonsLabour market shortages

• Women now make up over half of graduates, but only around 20% of graduates in STEMNeed to promote women in STEM

Definition of STEM: Persons with Bachelor degree or higher by field of study,HILDA 2016 (population estimates)

Field of study Persons Female

(%) Natural and physical sciences 264,540 46.5 Information technology 260,011 32.0 Engineering and related technologies 401,248 14.4 Architecture and building 100,820 29.0 Agriculture, environment and related studies 105,459 37.1 Medicine 157,939 46.4 Nursing 350,809 91.3 Other health-related 317,485 71.4 Education 765,976 71.4 Management and commerce 1,029,647 44.3 Law 173,621 43.8 Society and culture 630,985 68.9 Creative arts 186,242 58.9 Food, hospitality and personal services 43,738 48.9 Other 70,400 61.3 Total 4,858,920 54.3

STEM

STEM and the labour market• Labour market demand

• ShortagesHealy et al. (2016) – shortages apparent in engineering, employers report difficulties

filling vacancies• Over-supply

Norton (2016) - more STEM graduates than market can absorb

• For women?• Li et al. (2017), recent graduates earn 16% lower wages, markedly less likely

to be in a well-matched job• Barriers to promotion, lack of family-friendly work arrangements, lack of role-

models, high reported incidence of discrimination and (in US) sexual harassment.

10-year trends in STEM population in Australia:2006 and 2016 ABS Census

2006 2016

Male Female Total % Female Male Female Total % Female Natural & physical sciences 104,034 84,866 188,900 44.9% 146,587 138,283 284,870 48.5% Information technology 88,055 29,123 117,178 24.9% 161,741 57,365 219,106 26.2% Engineering & related Tech. 167,674 23,498 191,172 12.3% 285,577 54,954 340,531 16.1%

STEM 359,763 137,487 497,250 27.6% 593,905 250,602 844,507 29.7% Other fieldsa 786,979 1,198,060 1,985,039 60.4% 1,255,673 2,081,218 3,336,891 62.4%

All graduates 1,146,742 1,335,547 2,482,289 53.8% 1,849,578 2,331,820 4,181,398 55.8% STEM share all graduates 31.4% 10.3% 20.0% 32.1% 10.7% 20.2%

Trends in labour market status NPS IT ERT STEM

non-STEM

2006 Census Participation rate Men 82.6% 93.4% 86.1% 86.9% 87.0% Women 77.9% 83.6% 79.7% 79.4% 80.3% Unemployment rate Men 3.04% 4.18% 2.63% 3.15% 2.66% Women 3.38% 4.80% 4.91% 3.96% 2.53%

2016 Census

Participation rate Men 80.2% 93.4% 85.8% 86.2% 82.9% Women 76.4% 80.3% 78.5% 77.7% 78.6% Unemployment rate Men 3.63% 3.92% 3.16% 3.49% 4.70% Women 4.28% 6.22% 6.02% 5.06% 4.13%

The Data: Household, Income and Labour Dynamics in Australia Survey (HILDA)• Panel established in 2001, data available for Waves 1 to 16• Representative sample of private households, all persons

aged 15 and over surveyed• ≈13,000 responding individuals from 7,000 households per

year• In 2011 top-up sample of 2,153 households with 4,009 responding individuals

HILDA Survey (cont’d)• Wealth of information on labour market experiences,

background demographics, attitudinal data, etc.• Data on overall job satisfaction and satisfaction with

different aspects of your jobWith pay, security, the work itself, hours, flexibility.

• Special education modules included in 2012 and 2016Field of study of highest education only asked in those

modules in 2012 & 2016

Identifying people with STEM qualifications

2001 20162012

All new persons asked theirhighest level of qualification

In each subsequent year, all continuing personsasked if they completed a qualification in previous yearand, if so, the level of that qualification

tf. We know level of highest qualificationin every year (derived variable)

Deriving panel data (Bachelor degree or above)

201220112010

Individual’s field of qualand highest level known(eg. Graduate dip/certin Information Tech.)

Grad dip/cert or higher completed in last year?

Delete 2011 &prior observations

Yes

No

Grad dip/cert or higher completed in last year?

Delete 2010 &prior observations

Yes

No

Grad dip/cert or higher completed in last year?

Delete 2009 &prior observations

Yes

Check: Highest qualis Grad dip/cert?

Yes

No

Delete 2011 & prior observations

No

Check: Highest qualis Grad dip/cert?

Delete 2010 & prior observations

No

Yes

Graduates with field of study of highest qualification determined, HILDA sample Observations (persons) Female share (%)

Wave STEM Non-STEM Total STEM Non-STEM Total 1 (2001) 218 952 1170 22.5 62.1 54.7 2 (2002) 233 1022 1255 24.0 61.8 54.8 3 (2003) 258 1104 1362 25.2 61.8 54.8 4 (2004) 284 1202 1486 24.3 61.9 54.7 5 (2005) 307 1331 1638 24.4 62.0 54.9 6 (2006) 319 1370 1689 25.4 62.4 55.4 7 (2007) 355 1550 1905 25.6 62.6 55.7 8 (2008) 379 1673 2052 26.4 61.9 55.4 9 (2009) 417 1829 2246 26.6 61.5 55.0 10 (2010) 444 2000 2444 27.7 62.3 56.0 11 (2011) 660 2899 3559 28.8 62.5 56.2 12 (2012) 715 3275 3990 28.7 62.5 56.4 13 (2013) 685 3238 3923 28.8 62.2 56.3 14 (2014) 669 3313 3982 28.8 62.4 56.8 15 (2015) 698 3438 4136 29.2 62.7 57.1 16 (2016) 744 3656 4400 30.6 63.0 57.5 Pooled 7385 33852 41237 27.6 62.3 56.1

Multivariate panel model results: HILDA Waves 1-16 Sample Males and females Females only

Independent. var Female STEM Female*STEM STEM Model/Dep var. β P>z β P>z β P>z β P>z Binary logit Participationa n.a. n.a. n.a. -0.37* 0.08 Unemployment -0.28* 0.07 -0.03 0.87 0.38 0.30 0.33 0.28 OLS Real hourly wages -0.11*** 0.00 0.09*** 0.00 -0.09** 0.02 0.00 1.00 Ordered probit (Satisfaction with …) Emp. opportunities 0.00 0.90 -0.04 0.48 -0.18* 0.06 -0.21*** 0.00 Total pay 0.02 0.59 0.08* 0.10 0.09 0.32 0.18** 0.02 Job security 0.00 0.98 -0.05 0.39 -0.04 0.69 -0.10 0.23 The work itself 0.06* 0.07 0.03 0.60 -0.10 0.22 -0.09 0.19 Hours worked 0.05 0.17 0.09** 0.04 0.18** 0.04 0.24*** 0.00 Flexibility -0.02 0.62 0.12** 0.01 0.15 0.11 0.24*** 0.00 Job overall 0.06 0.10 0.05 0.28 0.00 0.96 0.03 0.70 (disagree/agree…) Uses skills/abilities 0.16*** 0.00 0.06 0.28 -0.34*** 0.00 -0.27*** 0.00

Notes: ***, * *and * indicate the estimated coefficient is significantly different from zero at the 1, 5 and 10 percent levels, respectively; a. sample restricted to persons aged 69 and under.

Analyses by sub-discipline• The lower participation of females with STEM qualifications applies primarily to

those with qualifications in the natural and physical sciences• The higher incidence of unemployment is driven by higher unemployment for

women with information technologyEstimates suggest a female with IT qualifications is 4 times more likely than a male with IT

qualifications to be unemployed• Women’s reduced satisfaction with employment opportunities applies across the

three fieldsMost pronounced for IT

• Women with engineering related qualifications earn 20% higher hourly wages than women with non-STEM qualifications it is these women that drive the higher pay-satisfaction among STEM qualified women

• The perceptions of under-utilisation of skills and abilities applies to women with qualifications in the natural and physical science and in IT it does not apply to women in engineering fields.

Concluding thoughts• Sceptical of arguments of the need for more STEM graduates• Policies to encourage more women into STEM need to be

more carefully considered:Supported with demand-side (workplace) policies?Compensatory incentives?

• More research needed to validate arguments and to guide policyDifferences by sub-disciplines seem important

Mixed evidence re. effects of male domination

• Significant limitations relating to definitions and data